In spite of all prevention efforts, 56,000 new HIV infections take place in the USA each year. Leaders in the field are calling for structural or social interventions that affect the risk environments, networks and social norms that influence risk behaviors and/or the probability that an HIV uninfected person will engage in sex or drug injection with an infected person. Developing structural interventions that work will be difficult and will require considerable research. As has been the case with HIV behavioral interventions, vaccine trials and microbicide research, initial structural intervention efforts are likely to meet only partial success (if that). In order to learn from these trials, it is essential to be able to measure the pathways that connect the target of the intervention-typically, a law, an organizational characteristic, or a social program such as housing provision- to changes in behaviors, risk environments or risk networks and then to STI or HIV incidence. This project aims to develop reliable and valid measures of important intermediate pathways between upstream changes or interventions and the immediate causes of infection. These pathways will include a. individuals' perceived immediate social environments; b. how individuals' lives are structured in ways that might affect how they react to changes in their environments; and c. how individuals react to normative influence from others in their environments (and the norms they develop as a result). The specific aims of the project are to develop these measures; to assess their reliability and concurrent criterion validity; and then to describe the distributions of these variables among reasonably representative samples of injection drug users, high-risk heterosexuals, and men who have sex with men in New York City (the epicenter of the US HIV epidemic). We will do this by using ethnographic approaches to develop the measures; and then by surveying 290 injection drug users, 290 high-risk heterosexuals, and 290 men who have sex with men to describe their psychometric properties and their population distributions. We will make these measures available to researchers in the USA and globally so they can use them - perhaps in modified culturally competent form for other countries - to develop and evaluate structural interventions; and, in an allied research area - to understand how large-scale social changes like wars or large-scale socioeconomic changes sometimes do and sometimes do not lead to HIV epidemics.